To the Editor In their interesting clinical article, Boige et al1 reported a statistically significant association between 2 DPYD variants (ie, D949V and V732I [DPYD*6]) and increase in severe toxic effects among patients with colon cancer treated with an adjuvant fluorouracil-based regimen. The issue of detecting dihydropyrimidine dehydrogenase (DPD) deficiency in patients scheduled for a fluorouracil- or capecitabine-based regimen is an old, critical, yet still ongoing story. Upfront detection could indeed help to tailor fluorouracil dosing, thus reducing the risk of treatment-related toxic effects in deficient patients. To date, several DPYD gene variants such as DPYD*2A, DPYD*5 (rs1801159t>C), DPYD*6 (rs1801160 C>T), DPYD*9A (rs1801265 A>G), and to a lesser extent DPYD*3 (rs72549303 C>del), DPYD*13 (rs55886062 A>C), and rs67376798T>A have all already been possibly associated with such increased risk.2 Of note, most gene-candidate studies have failed to evidence a sex effect. Yet, the fact that women are more prone than men to severe toxic effects when treated with fluorouracil or capecitabine has been known for several years and has been repeatedly evidenced.3,4 Interestingly, in the article by Boige et al,1 such a link between sex and risk of toxic effects was confirmed with women. Unfortunately, the authors did not provide further informations about a possible sex effect in the incidence of D949V and V732I polymorphisms in the studied population. Should a sex effect not be found, and because of the low sensitivity of gene-candidate studies, the clinical relevance of prospective DPYD genotyping as the backbone strategy to screen and to detect patients at risk could be questioned, despite its high specificity. To what extent screening for these single-nucleotide polymorphisms could reduce the high rate of severe toxic effects observed in women treated with fluorouracil remains therefore to be discussed or elucidated. Of note, several other approaches, including phenotyping, have been presented for years as possible complementary strategies to sort patients by DPD status. Recent reports suggest that simple DPD-driven adaptive dosing of fluorouracil could sharply reduce the incidence of severe toxic effects in patients with digestive cancer treated with a fluorouracil-containing regimen.5 Consequently, in addition to DPYD gene screening as proposed by Boige and colleagues,1 we believe that developing functional testing at the bedside could probably help to better picture the actual DPD status of patients scheduled for fluorouracil-based therapy, so as to propose subsequent appropriate adaptive dosing strategies. Thus, we advocate here for developing more comprehensive approaches (ie, combining genotyping and phenotyping strategies) for the upfront screening of DPD deficiency as a means to secure the administration of this widely prescribed drug, especially in women.